Abstract
Solvent-based heavy oil recovery methods are of interest as environmentally
friendly alternatives for thermal techniques. The phase behaviour data from a
heavy oil/solvent system are important information required for feasibility
studies and numerical simulation of such processes. The scarcity of
experimental data in the literature is a challenge in modelling of solvent
involving processes. The variety of the solvent/oil mixtures, which are being
evaluated within ongoing researches such as the VAPEX (vapour extraction of
heavy oil) process, requires accurate description of the system’s pressure,
volume and temperature (PVT) properties.
In this study, an experimental setup was designed to perform a series of PVT
experiments and viscosity measurements. The results of the PVT tests conducted
with the Frog Lake heavy oil and butane as a solvent are presented. The same
oil/solvent pair was used in the VAPEX experiments previously reported by the
authors(1, 2). The measurements include the solvent solubility in
the oil, mixture density and mixture viscosity at different saturation
pressures.
To simulate the phase behaviour of the system, an equation of state (EOS)
was tuned using the measured experimental data and a phase behaviour package
(WINPROP). The predicted densities and saturation pressures by the EOS are in
very good agreement with the experimental data. A mixing viscosity correlation
was also tuned with the measured data and found to be representative for
describing the viscosity of the system. The viscosity data were compared with
the predictions of several other available correlations, and it was shown that
Shu’s model(3) reproduces acceptable data for reservoir simulation
purposes.
Introduction
Solvent-involving recovery processes have recently gained some attention.
These processes often involve relatively light hydrocarbon solvents such as
C3 – C7, which are sometimes co-injected with
non-condensable gases such as CO2, CH4 and N2.
Numerical simulation studies of such processes are, however, in early stages to
investigate the feasibility of field implementation, improvement and
optimization. Numerical modelling of these processes is mostly performed on
compositional simulators to capture the potential compositional changes,
asphaltene precipitation and diffusion/dispersion mechanisms. Phase behaviour
of the heavy oil/solvent system is one of the most vital pieces of input data
that can be predicted and produced by either a series of k values or a
tuned EOS. Nonetheless, both methods rely on accurate experimental phase
behaviour information.
© 2010. Society of Petroleum Engineers
View full textPDF
(
848 KB
)
History
- Original manuscript received:
26 March 2007
- Meeting paper published:
12 June 2007
- Revised manuscript received:
8 August 2009
- Manuscript approved:
4 December 2009
View Cart
